Randomized Controlled Trial

A - Click here for explanation of classification scheme.

POSITIVE: See Quality Criteria Checklist below.

To determine whether homocysteine-lowering therapy reduces the risk of ischemic and hemorrhagic stroke, as well as stroke-related disability.

• Men and women 55 years of age or older, irrespective of their homocysteine levels from countries: 
  • With mandatory folate fortification of food (Canada and the United States)
  • Without mandatory folate fortification (Brazil, western Europe and Slovakia)
• History of:
  • Stroke
  • Transient ischemic attack (TIA) with objective evidence of ischemic ischemic cerebrovascular disease or with endarterectomy
  • Vascular disease (coronary, cerebrovascular or peripheral vascular)
  • Diabetes and additional risk factors for atherosclerosis.

Patients who were taking vitamin supplements containing more than 0.2mg of folic acid per day.

Recruitment
 

Subjects were recruited from 145 participating centers in 13 countries:

• Canada (N=3,568)
• United States (N=414)
• Brazil (N=265)
• Western Europe (N=426)
• Slovakia (N=849).

Design
 

Randomized clinical trial.
 

Blinding Used
 

Unclear as to whether the study with single or double blind. The discussion indicates that the study was "masked."
 

Intervention
 

• Participants were randomized (central randomization) to receive a once-daily supplement containing 2.5mg of folic acid, 50mg of vitamin B₆, and 1.0mg of vitamin B₁₂, or matching placebo, for five years
• No request was made to unmask treatment allocation for any given participant
• Participants were assessed in study clinics at six-month intervals.

Statistical Analysis

• The distribution of homocysteine was statistically significantly skewed so these data were log-transformed and inverse transformations were used to generate geometric mean values
• All analyses were by intention-to-treat, comparing the effect of homocysteine-lowering therapy against placebo. Binary stroke outcomes were analyzed by Cox proportional hazards regression models and risk estimates were expressed as a crude HR and 95% CI. Survival curves were estimated according to the Kaplan-Meier procedure and compared between treatment groups using a log-rank test. In the rare (less than 1%) circumstance that an individual could not be assessed at a clinic visit or contacted by telephone, the individual was considered to be free of a cerebrovascular event at that point in time. If lost to follow-up, then a participant was censored at the time of last contact. Censoring also occurred when a participant died or experienced any form of stroke, but not a TIA. Risk of stroke in pre-specified subgroups, including sex, those older and younger than the median age of the study participants, those with and without a history of hypertension, diabetes, smoking, coronary artery disease, TIA/stroke, cholesterol levels above or below the median of study participants, use of lipid-lowering therapy, use of antithrombotic agents, region (mandatory food fortification with folic acid vs. non-fortified areas), and quartiles of homocysteine levels was also assessed. In addition, a post-hoc analysis using a multivariable Cox proportional hazards models to evaluate the effect of homocysteine-lowering vitamin therapy on stroke adjusted for age, sex, and extended use of an antihypertensive, antiplatelet and lipid-lowering agents (model 1) and age, sex and geographic region (model 2) was conducted
• Among persons with an incident stroke, the net mean difference in change in neurological deficit from initial presentation to that at 24 hours was compared between homocysteine-lowering therapy and placebo groups using a Wilcoxon rank sum test. The trend in the modified Rankin scores at hospital discharge (or seven days after stroke onset) was assessed using a Mantel Haenzel 2 test. Modified Rankin scores were categorized into two levels of function: Independence (a score between zero and two) and dependence (a score of three to six). The risk of being functionally dependent after stroke was expressed as an OR and 95% CI, comparing persons who received homocysteine-lowering therapy to those assigned placebo.
• Statistical significance was set at a two-sided P<0.05 for all analyses, which were performed using SAS version 9.1 (SAS Institute Inc.).  

Timing of Measurements
 

• Measurements of study endpoints, side effects and adherence were repeated at six-month intervals
• Adherence to treatment was assessed by interview and pill count.
   

Dependent Variables

• Plasma homocysteine after overnight fast, by fluorescence polarization immunoassay (Abbott IMX; Abbott Laboratories, Inc.)
• All stroke events, but not TIA, were centrally adjudicated by the Adjudication Committee using pre-specified definitions of outcomes and all available supporting source documentation. TIA was excluded from all stroke outcomes. A disabling stroke was based on a modified Rankin score of three to five, equivalent to having moderate to severe disability. A stroke was defined as a focal neurological deficit lasting more than 24 hours. Computed tomography or magnetic resonance cranial imaging was obtained for 92% of participants with an ischemic stroke. Among persons with an incident stroke, the following symptoms were recorded at stroke onset and at 24 hours thereafter:
  • Change in the level of consciousness
  • Ocular or visual symptoms (diplopia or amaurosis fugax)
  • Motor weakness
  • Sensory symptoms
  • Dysarthria or dysphasia
  • Dysphagia
• A higher number of symptoms suggested a worse neurological deficit
• Separately, stroke-related disability at discharge from hospital or at seven days after stroke (whichever came first) was determined using the modified Rankin scale, graded from zero to six, as described in Supplementa.

Independent Variables
 

Number of cases of stroke.
 

Control Variables
 

• Baseline demographic data, medical history and medication use, including current use of anticoagulant therapy, were recorded for all participants at study entry. History of stroke or TIA was also documented
• After an overnight fast, blood was obtained for measurement of baseline plasma homocysteine from 3,306 randomly selected participants.

• Initial N: 5,522 (28% women)
• Attrition: 21 participants in the active arm and 16 in the placebo arm were lost to follow-up or withdrew from the study, but all were enrolled for at least two years and were included in the final analysis, censored for duration of observation
• Mean age: 69 years
• Ethnicity: Participants were from Canada, US, Brazil, Western Europe and Slovakia
• Other relevant demographics: 
  • There was no significant differences in baseline characteristics between those receiving placebo and therapy
  • The geometric mean homocysteine concentration at baseline, measured among 3,306 participants, was 11.5µmol per L in both groups. In the active group, the mean concentration was 2.1µmol per L lower in countries with folic acid food fortification than in those without fortification; in the placebo group, this difference at baseline was 2.3µmol per L. At the end of the trial, the mean homocysteine concentration was 9.3µmol per L in the vitamin therapy group and 12.3µmol per L in the placebo group, indicating a decrease of 2.2µmol per L and an increase of 0.80µmol per L, respectively.
• Anthropometrics:

Table 1. Baseline Characteristics of Study Participants

Baseline Characteristics*	Homocysteine-Lowering Therapy (N=2,758)	Placebo (N=2,764)
General characteristics
Mean (SD) age, year	68.8 (7.1)	68.9 (6.8)
Female	796 (28.9)	763 (27.6)
Ethnicity, non-white	106 (3.8)	109 (3.9)
Living in North America	1,988 (72.1)	1,994 (72.1)
Cardiovascular risk factors
Coronary artery disease	2,285 (82.8)	2,315 (83.8)
TIA	158 (5.7)	154 (5.6)
Stroke	241 (8.7)	251 (8.9)
TIA or stroke	341 (12.4)	343 (12.4)
Peripheral artery disease	216 (7.8)	169 (6.1)
Hypertension	1,542 (55.9)	1,497 (54.2)
Dyslipidemia	1,524 (55.3)	1,526 (55.2)
Diabetes mellitus	1,122 (40.7)	1,087 (39.3)
Current smoking	306 (11.1)	327 (11.8)
Baseline medication use
Antiplatelet agent	2,148 (77.9)	2,224 (80.5)
Oral anticoagulant	227 (8.2)	193 (7.0)
Lipid-lowering agent	1,627 (59.0)	1,690 (61.1)
Estrogen replacement therapy among women	137 (17.2)	130 (17.0)
Multivitamin supplement	331 (12.0)	307 (11.1)
Baseline physical measures
Mean (SD) heart rate, bpm	68.7 (11.2)	68.9 (11.5)
Mean (SD) systolic blood pressure, mmHg	138.8 (21.7)	138.9 (23.4)
Mean (SD) diastolic blood pressure, mmHg	77.4 (11.8)	77.5 (11.7)
Mean (SD) BMI, kg/m2	29.6 (16.4)	29.7 (21.1)
Baseline biochemical measures
Mean (SD) serum total cholesterol, mmol per L	4.83 (1.01)	4.78 (0.98)
Mean (SD) serum LDL cholesterol, mmol per L	2.73 (0.85)	2.68 (0.83)
Mean (SD) serum HDL cholesterol, mmol per L	1.21 (0.36)	1.18 (0.34)
Mean (SD) serum triglycerides, mmol per L	2.02 (1.35)	2.04 (1.29)
Mean (SD) serum glucose, mmol per L	7.15 (3.22)	6.97 (2.88)
Geometric mean (SD) plasma homocysteine concentration, µmol per L	11.5 (0.80)	11.5 (0.80)

*Data are presented as a number (percentage) unless otherwise indicated.

A total of 3,306 participants underwent measurement of plasma homocysteine at baseline.

HDL indicates high-density lipoprotein; LDL, low-density lipoprotein.

• Location: Individuals were recruited from 145 participating centers within 13 countries, including Canada (N=3,568), US (N=414), Brazil (N=265), Western Europe (N=426) and Slovakia (N=849).

• There were 258 (4.7%) strokes in total. 111 (4.0%) patients in the homocysteine-lowering group and 147 (5.3%) in the placebo group had a stroke (HR, 0.75; 95% CI, 0.59 to 0.97; absolute risk reduction, 1.3%; Table 2)
• The corresponding risk of ischemic stroke (HR, 0.81; 95% CI, 0.60 to 1.09) and hemorrhagic stroke (HR, 0.80; 95% CI, 0.32 to 2.02) tended to favor vitamin therapy, but not significantly so (Table 2)
• The relative risk of stroke was most reduced among those who baseline homocysteine concentration was in the highest quartile
• The effect of therapy was greater among subgroups of person younger than age 70 years, with untreated hyperlipidemia, those not receiving an antiplatelet agent and in persons with hyperhomocystinemia or residing in a country without folic acid food fortification.

Table 2. Stroke Outcomes Comparing Homocysteine-Lowering Therapy vs. Placebo

 
 

	N (Incidence Rate per 100 Person-Years)
Outcome*	Homocysteine-Lowering Therapy (N=2,758)	Placebo (N=2,764)	HR (95% CI) Homocysteine-Lowering Therapy vs Placebo
Any stroke	111 (0.88)	147 (1.15)	0.75 (0.59–0.97)
Ischemic stroke	79 (0.62)	98 (0.77)	0.81 (0.60–1.09)
Hemorrhagic stroke	8 (0.06)	10 (0.08)	0.80 (0.32–2.03)
Non-fatal stroke	84 (0.66)	117 (0.92)	0.72 (0.54–0.95)
Fatal stroke	27 (0.21)	30 (0.24)	0.91 (0.54–1.53)
Disabling stroke	26 (0.21)	41 (0.32)	0.64 (0.39–1.04)
Fatal or disabling stroke	48 (0.38)	63 (0.49)	0.77 (0.53–1.12)

*If a participant had more than one outcome during the study period, only the first event was included in this analysis.

Table 3. Change in Neurological Deficit at the Onset of Stroke Symptoms and at 24 Hours Thereafter, Comparing Homocysteine-Lowering Therapy vs Placebo

	Mean (SD) Neurological Deficit Score*
	At Initial Onset of Stroke Symptoms		At 24 Hours After Onset of Stroke Symptoms		Change Between 24-Hour and Stroke Symptoms at Initial Onset
Stroke Type	Homocysteine-Lowering Therapy	Placebo	Homocysteine-Lowering Therapy	Placebo	Homocysteine-Lowering Therapy	Placebo	P
Any stroke	2.2 (1.4)	2.3 (1.5)	1.4 (1.4)	1.5 (1.4)	–0.85 (1.4)	-0.77 (1.4)	0.45
Ischemic	2.4 (1.3)	2.5 (1.5)	1.4 (1.2)	1.6 (1.4)	–0.96 (1.4)	-0.93 (1.5)	0.69
Hemorrhagic	1.6 (1.9)	2.2 (1.6)	1.4 (1.9)	1.8 (1.3)	–0.25 (0.89)	-0.4 (1.3)	1.00

 
 

*Defined by the number of symptoms as follows:

• Change in level of consciousness
• Ocular or visual symptoms
• Motor weakness
• Sensory symptoms
• Dysarthria or dysphasia
• Dysphagia.

A higher score indicates a greater neurological deficit.

Wilcoxon rank sum test comparing the net mean difference in the change in the neurological deficit score between homocysteine-lowering therapy vs. placebo.

• In conclusion, combined daily administration of 2.5mg of folic acid, 50mg of vitamin B₆, and 1.0mg of vitamin B₁₂ for five years had a modest, but beneficial effect on stroke prevention or fatal stroke among a population at high risk for cardiovascular disease
• Ongoing clinical trials should help determine whether folic acid and B vitamin supplements are efficacious in reducing stroke risk and severity and which patient subsets are most likely to derive benefit from this treatment.

The author's noted the following limitations:

• Only approximately 12% of participants had a historyof stroke or TIA. Nevertheless, the trial a fairly large number of stroke events
• Also, because the mechanisms for ischemic stroke were not captured, they cannot comment whether homocysteine-lowering therapy may differentially impact on large vs. small artery disease
• The non-significant reduction in disability at seven days or at discharge is not surprising, because stroke recovery often requires weeks rather than days, and this would lower the ability to detect a clinically important difference, especially given the statistical power of only 30% for ischemic stroke and 42% for disabling stroke in the current trial
• The original HOPE-2 trial was designed and powered to detect a proportional reduction in the risk of cardiovascular death, myocardial infarction and stroke, rather than just stroke or one of its subtypes
• Approximately two-thirdsof the study patients were recruited from countries with folate food fortification, which may impact on the ability of the study intervention to further lower homocysteine levels. However, compared to the placebo group, homocysteine levels at study end were approximately 25% lower in the active treatment group
• The results of this secondary analysis of the HOPE-2 trial may only be applicable to adults older than 55 years with atherosclerosis.